Message12332

Jeff, thanks for digging this up! It seems to me that such a development loops occur regularly in Jython, probably in other sufficiently complex projects too. A central question popping up is: Why did testing not prevent reintroduction of old bugs?
The answer is obvious: Fixing the issue had much higher priority than writing a test. Often enough a proper test is difficult to implement. I also found myself unable to allocate additional time to proper test-writing in the phase before 2.7.1 release. This is really a symptom of lacking resources.

Nevertheless we should reason about these lessons learned and consider ways to prevent this in the future. Maybe this story should somehow go into developer guide..., at least as a link?
I imagine it was some work to dig this up and it shouldn't be lost (i.e. de-facto invisible once this issue gets closed).